First Advisor

Lalonde, Trent L.

Document Type

Dissertation

Date Created

8-2017

Department

College of Education and Behavioral Sciences, Applied Statistics and Research Methods, ASRM Student Work

Abstract

Longitudinal data occur in different fields such as biomedical and health studies, education, engineering, and social studies. Planning advantageous research projects with both high power and minimum sample size is an important step in any study. The extensive use of longitudinal data in different fields and the importance of their power estimation, yet the limited resources about their respective power estimation tools, made it worthwhile to study their power estimation techniques. The presence of time-dependent covariates triggers the need to use more efficient models such as generalized method of moments than the existing models which are based on generalized estimating equations. Not taking into consideration the correlation among observations and the covariates that change over time while calculating power and minimum sample size will cause expensive research being conducted without using data that are capable of answering the research questions (Williams, 1995). Two different power estimation and minimum sample size calculation techniques for longitudinal data in the presence of time-dependent covariate using generalized method of moments approaches are constructed in this study and their performances are evaluated.

Keywords

Time-series analysis; Longitudinal method; Linear models (Statistics)

Extent

166 pages

Local Identifiers

Ramezani_unco_01610D_10596

Rights Statement

Copyright is held by the author.

Share

COinS